206 ALASKAN GLACIEK STUDIES
the moraine may be from 5 to 10 feet deep. Marginal bands of deeper moraine also extend up the lateral margins of the valley glaciers, marking the site of the lateral moraine accumulation. There are crescentic bands of thicker moraine in the piedmont bulbs, giving rise to ridges because of retarded ablation, while crescentic valleys lie between the ridges where the moraine is thinner. As ablation proceeds the position of the ridges and valleys change, for when the ridges become too steep the moraine slides into the depressions. Speaking generally, there is a gradational decrease in thickness of ablation moraine from the periphery of the piedmont bulb to the part of the valley glacier where the ablation moraine disappears, a point which varies greatly from glacier to glacier, but which is always in the area of thedissipator, and well below the snow line. This decrease in thickness is locally interrupted by areas of thickening on ridge tops and of thinning in valleys and on steep ridge slopes.
There are variations in nature of material composing the ablation moraine. Speaking generally it is made up of frost-riven angular fragments of rock of the kind enclosing the upper glacier, and often includes bowlders of huge size. Scratched stones are also found, though not commonly, and there are even areas of clay and waterwashed gravel, for short streams flow on the moraine, and pools are not uncommon. There are bands* often crescentic, in which rocks of one kind so predominate as to give rise to bands of color on the glacier surface, as in the Variegated Glacier. With our lack of knowledge of the behavior of ice currents in spreading glaciers, and our ignorance of the bed rock in most of the enclosing valley walls, it is not possible to offer a definite explanation of these variations. They are without question due to the flow of the ice in the piedmont bulb, distributing the load which the ice bears and which is brought into prominence by ablation; but the nature of the process is not yet clear to us.
On the thicker, outer portion of the ablation moraine the soil has such stability that vegetation grows luxuriantly and one often needs to study closely to determine where the land forest ends and the glacier forest begins. It seems that from 5 to 10 feet of moraine on ice in this climate is sufficient to give rise to such a condition of stability as to permit practically uninterrupted forest growth with spruce, hemlock and cottonwood trees like those on land. The density and maturity of the forest growth progressively diminishes from the peripheral zone, and near its inner margin there are abundant signs of the struggle to which plants are subjected when growing in a slumping soil. The average age of the plants, here, mainly, if not entirely, alders and willows, may be from ten to twenty years, and the majority of individuals are healthy and undisturbed. But areas occur where for some reason ablation is locally more active, and there plants are found overturned, others with their roots partly uncovered, others partly inclined; in fact all stages in plant destruction are to be seen, and perhaps on the slopes of some basin containing a pool, or where a moulin has opened in the moraine, the ice itself may even be seen. Looking down upon a forest-covered ablation moraine from, some favorable viewpoint, numerous areas of this sort may be seen. They seem to be related to the-development of ice drainage in the vegetation-covered parts of the piedmont bulb; but they occupy only a small proportion of the vegetation-covered part of the moraine and are mainly confined to the inner part of the zone. Even here, however, the condition of soil sjaibility is in general sufficient to permit the growth of mature alders.
From the zone of occasional slumping there is a rapid gradation to the barren zone of the ablation moraine, which is a true desert, almost devoid of life. In this gradational